# CATCH: A Controllable Theme Detection Framework with Contextualized Clustering and Hierarchical Generation

arXiv: https://arxiv.org/abs/2512.21715 · HTML: https://mlsanity.com/paper/2512.21715/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 36 references that are in the map.
- First posted on arXiv 25 Dec 2025.

## Builds on (most cited first, 23 of 23)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [BERTopic: Neural topic modeling with a class-based TF-IDF procedure](https://mlsanity.com/paper/2203.05794/) (2022)
- [Tired of Topic Models? Clusters of Pretrained Word Embeddings Make for Fast and Good Topics too!](https://mlsanity.com/paper/2004.14914/) (2020)
- [ClusterLLM: Large Language Models as a Guide for Text Clustering](https://mlsanity.com/paper/2305.14871/) (2023)
- [Is Neural Topic Modelling Better than Clustering? An Empirical Study on Clustering with Contextual Embeddings for Topics](https://mlsanity.com/paper/2204.09874/) (2022)
- [Improving Unsupervised Dialogue Topic Segmentation with Utterance-Pair Coherence Scoring](https://mlsanity.com/paper/2106.06719/) (2021)
- [Intent Mining from past conversations for conversational agent](https://mlsanity.com/paper/2005.11014/) (2020)
- [Uncovering the Potential of ChatGPT for Discourse Analysis in Dialogue: An Empirical Study](https://mlsanity.com/paper/2305.08391/) (2023)
- [Dialog Intent Induction with Deep Multi-View Clustering](https://mlsanity.com/paper/1908.11487/) (2019)
- [Automatic Intent-Slot Induction for Dialogue Systems](https://mlsanity.com/paper/2103.08886/) (2021)
- [Unsupervised Dialogue Topic Segmentation with Topic-aware Utterance Representation](https://mlsanity.com/paper/2305.02747/) (2023)
- [Intent Induction from Conversations for Task-Oriented Dialogue Track at DSTC 11](https://mlsanity.com/paper/2304.12982/) (2023)
- [Open Intent Discovery through Unsupervised Semantic Clustering and Dependency Parsing](https://mlsanity.com/paper/2104.12114/) (2021)
- [NeuroMax: Enhancing Neural Topic Modeling via Maximizing Mutual Information and Group Topic Regularization](https://mlsanity.com/paper/2409.19749/) (2024)
- [Bag of biterms modeling for short texts](https://mlsanity.com/paper/2003.11948/) (2020)
- [GloCOM: A Short Text Neural Topic Model via Global Clustering Context](https://mlsanity.com/paper/2412.00525/) (2024)
- [Dynamic transformation of prior knowledge into Bayesian models for data streams](https://mlsanity.com/paper/2003.06123/) (2020)
- [Advancing Topic Segmentation and Outline Generation in Chinese Texts: The Paragraph-level Topic Representation, Corpus, and Benchmark](https://mlsanity.com/paper/2305.14790/) (2023)
- [Unsupervised Mutual Learning of Discourse Parsing and Topic Segmentation in Dialogue](https://mlsanity.com/paper/2405.19799/) (2024)
- [Dialog Intent Induction via Density-based Deep Clustering Ensemble](https://mlsanity.com/paper/2201.06731/) (2022)
- [Towards a Fully Unsupervised Framework for Intent Induction in Customer Support Dialogues](https://mlsanity.com/paper/2307.15410/) (2023)
- [Overview of Dialog System Evaluation Track: Dimensionality, Language, Culture and Safety at DSTC 12](https://mlsanity.com/paper/2509.13569/) (2025)

## Abstract (arXiv)

Theme detection is a fundamental task in user-centric dialogue systems, aiming to identify the latent topic of each utterance without relying on predefined schemas. Unlike intent induction, which operates within fixed label spaces, theme detection requires cross-dialogue consistency and alignment with personalized user preferences, posing significant challenges. Existing methods often struggle with sparse, short utterances for accurate topic representation and fail to capture user-level thematic preferences across dialogues. To address these challenges, we propose CATCH (Controllable Theme Detection with Contextualized Clustering and Hierarchical Generation), a unified framework that integrates three core components: (1) context-aware topic representation, which enriches utterance-level semantics using surrounding topic segments; (2) preference-guided topic clustering, which jointly models semantic proximity and personalized feedback to align themes across dialogue; and (3) a hierarchical theme generation mechanism designed to suppress noise and produce robust, coherent topic labels. Experiments on a multi-domain customer dialogue benchmark (DSTC-12) demonstrate the effectiveness of CATCH with 8B LLM in both theme clustering and topic generation quality.